IE 11 is not supported. For an optimal experience visit our site on another browser.

Tweaks to BCS nearly complete

Formula to rely more on humans, less on computers
/ Source: a href="" linktype="External" resizable="true" status="true" scrollbars="true">The Washington Post</a

When last year's controversy surrounding the Bowl Championship Series reached a fever pitch, it became clear that the formula used to determine the teams that play for college football's national championship would be tweaked.

Now, the tweaks are nearly in place -- with no guarantees the old problems will be eliminated.

Representatives of the six conferences that make up the BCS, as well as the commissioners of the remaining five Division I-A conferences, met in Chicago yesterday to complete changes to the formula. Two sources familiar with the discussions, who requested anonymity because the plan had not been formally approved, said the new equation would be simpler, taking into account only the Associated Press writers' poll, the USA Today/ESPN coaches' poll and a combination of computer polls. Eliminated as factors would be a strength of schedule rating, a penalty for the number of losses and a bonus for wins over other teams that finish high in the BCS standings.

"I think it will change for this fall," ACC Commissioner John Swofford said yesterday. "I think it will be more simple, and I think it will have a stronger human element."

The new formula specifically addresses the controversy that erupted last year, when Southern California was ranked first in both the coaches' and writers' polls, but came out third in the BCS standings at the end of the regular season. That allowed Oklahoma, third in both human polls, to play Louisiana State in the Sugar Bowl for the national title despite the fact that the Sooners were beaten soundly by Kansas State in the Big 12 title game.

Had the new formula been in place last season, USC would have played LSU for the national title. Instead, the Trojans beat Michigan in the Rose Bowl and were crowned national champions by the AP, while LSU beat Oklahoma in the Sugar Bowl and won the official BCS championship.

Essentially, the new formula ensures that the computers won't over-ride the human voters. Should a team finish the regular season ranked No. 1 in both the writers' and the coaches' polls, it would be guaranteed a spot in the BCS title game, which this season will be the Orange Bowl.

But it is not infallible.

"Every time number one and number two in the polls don't meet, they feel they need to change it so that they would have," said Jerry Palm, who runs the Web site and charts the BCS. "But this doesn't prevent what happened last year from happening again. This isn't better. It's just different. . . . They think backwards, not forwards."

Palm said that while the new formula will be easier for fans to understand, it increases the power of both coaches and writers, who traditionally drop teams in the polls when they lose late in the season. Either coaches or writers could have other agendas, Palm said.

"Fans should be uncomfortable with the fact that it puts more influence into the hands of the voters," Palm said.

The commissioners and athletic directors in Chicago will also discuss further overhauls of the entire BCS system, including the idea of a "plus-one" plan that would involve the top two teams after the BCS bowls advancing to a title game a week later. Both conference commissioners and athletic directors, however, have said repeatedly that they may have a tough time selling that plan to college presidents, who are generally against extending the season.

Another concept being discussed would be to play the BCS bowls as they are and then play the national championship game a week later. Under that plan, one of the four BCS bowls -- the Fiesta, Orange, Rose or Sugar -- would host a second game each year.

"That would be a format that we would have extreme interest in," said Loren Matthews, the senior vice president for programming of ABC Sports.

Said Swofford: "I do think that would be one model that would be acceptable. I don't see that being problematic."